Hierarchical Natural Language Extraction for Database Generation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The conversion of unstructured natural language text information into structured database values is tedious and difficult due to variations in terminology, document structure, and unconventional terminology, leading to failed or incorrect interpretations.
Innovation Solution
Employing a sequence of deep learning models, including a multi-channel transformer model and finer-grain models, to extract and structure information from unstructured text documents, using hierarchical data extraction operations and micro-ai-models to enhance training efficiency and accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional natural language processing techniques are used to convert unstructured text to structured data, then the process becomes tedious and difficult, but the system lacks the capability to handle terminology variations and unconventional structures
Solution Approach 1:
The patent segments the text processing task into multiple hierarchical levels: document-level classification, section-level analysis, and entity-level extraction. This segmentation allows the system to handle complex terminology variations at different granularities, improving reliability without overwhelming system complexity
Solution Approach 2:
The patent introduces multi-dimensional processing by analyzing text across multiple dimensions: semantic meaning, structural context, and terminological variations. This dimensional approach enables the system to accurately interpret unconventional structures and terminology by examining them from multiple perspectives simultaneously
2Measurement precision
If hierarchical data extraction operations are implemented to improve accuracy, then data extraction accuracy improves, but the processing time and computational resources increase
Solution Approach 1:
The patent performs preliminary document classification and structure identification before detailed entity extraction. By pre-processing the document to establish its hierarchical structure and identify key sections, the system reduces the computational burden during the actual extraction phase, maintaining high precision while reducing processing time
Solution Approach 2:
The system uses the document's own structural cues and content patterns to guide the extraction process. By learning from the document's inherent hierarchy and terminology usage, the model adapts to each document type automatically, reducing the need for extensive manual configuration and reprocessing
Data Source
AI summary
Some embodiments may perform operations of a process that includes obtaining a natural language text document and use a machine learning model to generate a set of attributes based on a set of machine-learning-model-generated classifications in the document. The process may include performing hierarchical data extraction operations to populate the attributes, where different machine learning models may be used in sequence. The process may include using a pre-trained Bidirectional Encoder Representations from Transformers (BERT) model augmented with a pooling operation to determine a BERT output via a multi-channel transformer model to generate vectors on a per-sentence level or other per-text-section level. The process may include using a finer-grain model to extract quantitative or categorical values of interest, where the context of the per-sentence level may be retained for the finer-grain model.